Ont Epis Meth

7/29/2019 Ont Epis Meth

1/17

The University of Liverpool

Department of Social and Environmental Studies

School of Politics and Communication

Class: POLI 116 Research Methods in Politics

Lecturer: Dr. David Richards

Essay:

The Research Project

Student: Felix Poetschke

Student ID: 200153008

Course: MA Politics and Mass Media

Deadline: 01/05/2003


2/17

When embarking on a masters dissertation various points and problems have to be

considered. From the general stance towards the world to the specific problems of the

research method, it all influences the ultimate success of the project. However, this is not only

the case at a students project, but also at every research conducted by professional scholars.

This essay shall deal with general ideas regarding research but also with the particular

problems concerning my dissertation project.

Ontology Epistemology Methodology

The foundations on which political scientists work are their ontological and epistemological

positions. They are not always spelt out and are rather implicit than explicit, but show

themselves in the matter of methodology and approach. To Marsh and Furlong (2002) these

stances a political scientist takes are pivotal to his research, as they shape the approach to

theory and the methods utilised; and they are grounded deeply in the researchers beliefs

about the world, resulting in the effect that the positions taken on these issues cannot possibly

be changed: They are like a skin not a sweater: they cannot be put on or taken off whenever

the researcher sees fit.1 We will see further down if this argument could not be challenged.

Ontology is the science or theory of being. It concerns the question of how the world is built:

is there a real world out there that is independent of our knowledge of it?2 Put into the

political context the question might be What is the nature of the social and political contest

we might acquire knowledge about?3 Two basic distinctions can be made here: firstly, there

is a real (without quotation marks) world that is independent from our knowledge and upon

these foundations life is built hence the expressionfoundationalism ; or, secondly, there is

no real world but the world is socially and discursively constructed and hence dependent from

a particular time or culture.

Epistemology then is the theory of knowledge. Ones epistemological position reflects the

view of what we can know about the world and how we can know it.4 Again there are two

major distinctions to be made here: Firstly, it is possible to acquire knowledge about the

1 Marsh & Furlong 2002, p. 17.2

Marsh & Furlong 2002, p. 18.3

Hay 2002, p. 61.4

Marsh & Furlong 2002, p. 18-9.

2


3/17

world unmediated and with no interferences. This implies that objectivity is possible, because

everyone observes things in the same way. Secondly, observation is never objective but

always affected by the social constructions of reality. Obviously this relates back to

ontology. Foundationalists would take the former point of view, while anti-foundationalists

would employ the latter. For them there is no real world to observe, as every things or actions

obtain meaning only by actors and not by sheer existence. For the researcher this leads to the

problem of the double hermeneutic: the world is interpreted by actors and their

interpretation is interpreted by the observer5, making it a double interpretation even less

objective than the initial one.6

To summarize, there are two completely opposite positions with regard to ontology and

epistemology that have absolutely nothing in common. These are reflected in different

research traditions, to which I will turn now.

Positivism adopts a foundationalist ontology and an according epistemology. It developed

from the empiricist tradition of natural science and sees social science capable of the same

possibilities that are there in the natural science. That is, it is possible to observe everything

that happens and understand it as such without any mediation, thereby denying any

appearance/reality dichotomy. As in natural science theory is used to generate hypothesis,

which can simply be tested by way of direct observation. The ultimate aim is to find general

laws and causal statements about social phenomena. This implies that objectivity is possible.

Positivists usually use quantitative methods as research tools, as these are objective and the

results generalizable and replicable.7 They look for explanation of behaviour, not for the

meaning.

The opposite position is taken by relativists, also called interpretists. For them it is not

possible to make objective statement about the real world because there is no such thing as a

real world but it is only socially and discursively constructed. The ontological position here is

clearly anti-foundationalist. Because the world is only socially constructed so are social

phenomena, which positivists claim to be able to examine by sheer observation. This is not

possible, interpretists say, because they do not exist independently of our interpretation and

every observation concomitantly affects what we observe. Of course interpretist researchers

5Marsh & Furlong 2002, p. 19.

6See also Schmidt (1994a, 1994b), who distinguishes between observer of first and second order.

7Marsh & Furlong 2002.

3


4/17

also operate within discourses or traditions. Consequently, knowledge is theoretically or

discursively laden.8 Again the problem of the double hermeneutic has to be taken into

consideration here. Suiting the claims of not possible objectivity, relativists usually employ

qualitative research methods. Unlike positivists they look to understand social behaviour

rather than explain it and focus on its meaning.

However, there is a position that lies somehow in-between. Realists share positions from both

sides and form sort of a golden mean. They claim that there is a real world out there (so they

are foundationalists) and that it is possible to make causal statements. However, not all social

phenomena, and the relationships between them, are directly observable. There are deep

structures that cannot be observed and what can be observed may offer a false picture of those

phenomena/structures and their effects.9 So realism combines elements from both positivism

and interpretism.

Obviously all of these approaches are subject to diverse criticism. I will deal with each of

them in turn.

Positivism has come under attack from two different sides. The first concerns the problems

with objectivity and absolute reality. Objectivity is only then possible, when there is no

mediating factor that skews or alters the observation. But this, as Hollis and Smith (1990)

show, employing Quines argumentation, is not the case because the five senses do not and

cannot give us unvarnished news information independent of the concepts used to classify

it.10 We automatically use concepts to describe observations and these concepts inevitably

shape the outcome it is an interpretation rather then a pure observation. This means also that

when a theory is being tested, the theory will also affect the outcome of the observation,

because the theory is shaping the way we look at the observation and at the outcomes. There

can, therefore, be no objective observation separate from the theory.11 Another criticism

concerns the presumed parallels between social science and natural science. Critics argue that

there are fundamental differences between events in the natural and the social environment.

Social structures are shaped only by the constituting activities, and do not exist independently.

Secondly, the views of the agents acting in these social structures about them shape these

8 Marsh & Furlong 2002, p. 26.9

Marsh & Furlong 2002, p. 30.10

Hollis & Smith 1990, p. 55.11

Hollis & Smith 1990.

4


5/17

structures. As these views can change, the structures change also and can therefore vary

across space and time.12

The criticism of relativism is, obviously, directed in the exact opposite direction. It is about

the problem of validity and subjectivity: To positivists, the interpretist tradition merely offers

opinions of subjective judgements about the world. As such, there is no basis on which to

judge the validity of their knowledge claims. One persons view of the world, and of the

relationship between social phenomena within it, is as good as anothers view.13 This is only

a problem for positivists, as with their different ontological and epistemological view of the

world a different objective is given. However, also interpretists have tried to gain a certain

amount of objectivity.14 As Marsh and Furlong (2002) find in the work of Bevir and Rhodes,

a particular research or field of study is formed and influenced by historically produced

norms, rules and conventions, while the content has a certain narrative, that gives meaning to

the studies. Simplified, these traditions provide shared criteria, with which it is possible to

judge an argument true or false and an action right or wrong. It is therefore possible to assess

truthfulness of research. This corresponds with Schmidts (1994b) findings, who claims:

When producing sense or meaning, the most relevant aspects are those of collective

knowledge. These aspects aresharedby individuals (via rules, conventions, norms, common

sense) and via expected expectations enable social acting as well as are being confirmed in it

(knowledge about knowledge).15 In short, historical and social paradigms in research enable

the assessment of truth.

Realism, which shares positions of both interpretism and positivism, has to fight with

criticism from both sides. While positivists disagree with the notion of unobservable

structures, relativists cannot come to terms with the foundational claims of realism. However,

as most of the research, realism has turned more in the interpretists direction and has used

their criticism to adapt their position. Hence they acknowledge that interpretation of social

phenomena is crucial, and that differences between external and constructed reality have to be

identified and understood to explain social relationships.16

12Hay 2002.

13 Marsh & Furlong 2002, p. 27.14 The notion of an amount of objectivity may sound weird as the common knowledge might be that there is

either objectivity, and therefore a truth, or there is not. But if complete truth as a concept is rejected, there is stillthe chance of finding an approximation to reality.15

Schmidt 1994b, p. 615. Authors translation. Original emphasis.16

Marsh & Furlong 2002.

5


6/17

Now what can we make of all this and what does it mean? We have already seen that there is

the view that ontology and epistemology are the foundations on which a researcher must build

his research as they shape the approach to theory and the methods. Also it has been said that

the positions researchers take in these matters are like a skin not a sweater: they cannot be

put on or taken off whenever the researcher sees fit.17 There are two assumptions here:

Firstly, the method of our research is inevitably linked to our ontological and epistemological

position. Secondly, as these positions are not changeable, neither are the methods. This is

because, so Marsh and Furlong, these positions reflect fundamental views about the world,

which can be completely adversative, and a change in methods reflected a change in the

worldview, which is not possible for Marsh and Furlong. These claims pose a question: Is the

relationship between the ontology, epistemology and methodology really as directional as

described? Or, is there any way that a clear stance on these issues makes it possible to employ

different methods?

However, before we can answer these questions we must first establish what different

methodologies there are, what strengths and weaknesses each one has, and what the

relationship to the ontological and epistemological position is.

Basically there are two main methodological positions that are fighting for the researchers

attention. Qualitative and quantitative methods do share most of the same letters but otherwise

do not have much in common.

Quantitative methods are mostly employed by positivists. As they try to produce causal

explanations or even scientific laws they not only refer to the notion of natural science in their

ontology and epistemology, but also employ the same methods. In the end the methods

always result in numbers, which are then analysed for a proper result. The aim is to have no

interpretation in the analysis but to have direct and exact causations which are irrefutable. The

great advantages of this approach are that the data is usually easy to replicate, which is also a

very important factor for scientificness in natural science, and, especially, they are easy to

generalize. Typical methods of quantitative research are surveys or statistics. However,

corresponding with the general criticisms of positivism, the problem of this approach is that it

is never clear what the answers, in polls, for example, actually mean. It is of course the

question if this is a proper criticism, as the focus here is not on the meaning of behaviour but

17Marsh & Furlong 2002, p. 17.

6


7/17

on the explanation and causes of the behaviour. However, it is clear that even though this is a

very scientific approach, the notion of objectivity is no longer valid even here. The notion

of positivist objectivity criteria is by now generally accepted.18

Qualitative methods, on the other hand, are usually employed by relativists. Corresponding to

their ontological and epistemological position of a world that is only socially constructed and

all knowledge that we can have about it is subject to interpretation, relativists use interviews,

focus groups and other qualitative methods to get an in-depth sight into a field; with a

richness of description not obtainable by quantitative research. The aim is to find out the

meaning of social behaviour. While the richness of information cannot be disputed,

qualitative researchers have to face the problem that their work is hard to measure in terms of

reliability, validity and generalizability. Indeed there have been no real ways out of this

dilemma. Gavin, for example, notes in his account of focus group research: the issue of

generalization is still outstanding.19 However, there have been attempts to solve this

problem. Lunt and Livingstone (1996), for instance, claim that this critique simply does not

apply to qualitative methods: the notions of reliability and validity are inextricably linked

to quantitative methods and so are irrelevant to qualitative work.20 One could, in response,

argue that the notions of social communication and polysemic and context-dependent nature

of meaning21 are in turn inextricably linked to qualitative research and would therefore be

irrelevant to quantitative work. What argumentation one endorses here is probably depending

on ones ontological and epistemological position, although both of them miss the point.

Problems of both methodologies have to be acknowledged by researchers and none of them

can be treated as the sorcerers stone.

But where does this leave us with Marsh and Furlongs claim that the ontological and

epistemological positions are a skin rather than a sweater? When they say that these positions

can not be changed whenever the researchers sees fit22 then this can only mean that they

refer to the methodological implications of these positions and that researchers would claim

different positions to justify their chosen method. However, as alluded before, there have been

influences from the other positions so that hardliner positivist or interpretist researchers

hardly exist anymore, although the influence has been more from the interpretist corner. Also,

18 Hansen et al. 1998, p. 95.19 Gavin 1998, p. 172.20

Lunt & Livingstone 1996, p. 90.21

Lunt & Livingstone 1996, p. 90.22

Marsh & Furlong 2002, p. 17.

7


8/17

as Read and Marsh (2002) say, the differences between qualitative and quantitative methods

do exist, but they can easily be overstated.23 They refer to Bryman, who states that there is

nothing inherent in the properties of the different methodologies which prevents their use by

researchers who are operating from different epistemological positions.24 So researchers can

use the accordingly other method to correct short-comings of their preferred method. Surveys

might be constructed by interviewing small numbers of people first and testing question

wording or sequence. And it might be useful to cross-check interviews via content analysis on

possible incoherence in the findings. Also it seems imaginable that an interpretist researcher

has a problem to which the answer can best be found employing quantitative methods and

vice versa. On the other hand, then, maybe this seems not possible because exactly the

epistemology of the researcher prohibits these particular questions. However, while Marsh

and Furlong see a clear dependence between epistemology and methodology and Hay even a

directional dependence25, Read and Marsh find that the link between epistemology and

methodology is important, but far from determinant.26 To pick up Marshs and Furlongs

metaphor, it might not be a (woollen) sweater, and not a (human) skin, but perhaps a

snakeskin.

23 Read & Marsh 2002, p. 232.24

Read & Marsh 2002, p. 232-3.25

Hay 2002, p. 64.26

Read & Marsh 2002, p. 235.

8


9/17

The Research Project: Methods Problems Choices

This second part of this essay will deal with my particular research question for the

dissertation and all the possible problems and hurdles that I might encounter when trying to

answer it. My research will include the following: political advertising television spots of the

major parties in Germanys national elections of 1998 and 2002 will be analysed and the

extent to which they are negative will be assessed.

As we have seen before, a researchers ontology and epistemology shapes his/her

methodology. In reverse then there is also the possibility of inferring from the question

backwards to the ontology and epistemology. So what can be inferred from this particular

question? Firstly, the analysis will concentrate on thespots themselves; and secondly, the

extent, oramountof negativity shall be assessed. This leads us to two conclusions: As the

focus is on the spots, or so to speak a manifestation or real entity the researchers ontology

seems to be foundationalist. If this were not the case he would have asked to examine social

construction of things or similar. The second conclusion leads us to the epistemology: if the

researcher wants to measure an amount of something, in this case of negativity, then he must

deem that possible at all. This implies a rather positivist epistemology: it is possible to acquire

knowledge about the world unmediated and with no interferences. In view of these two

conclusions, the researcher would therefore though only measured on this one particular

question be in the positivist tradition rather than in the interpretist.

Thus, we can also predict how scholars from other traditions would have approached this field

of study. As mentioned above, the interpretist researcher would not have embarked on

examining the spots themselves, but would have tried to find out how people construct

meaning of these spots socially. Also, the question would not have been how much

negativity is in the spots, but how negative these spots are perceived by viewers. A typical

method for answering this would be the focus group.

However, when having thought of a research question, the obvious next problem is how to

answer it, that is, which method to employ. To a certain degree the question itself determines

how to answer it. If the amount of negativity is to be assessed, that is, how much of the

content of the spots is of negative nature, the most natural method of choice will be content

analysis. This will also be the method I am going to use. However, before outlining the

9


10/17

reasons for using content analysis and identifying its advantages and pitfalls, other possible

methods have to be considered.

Trying to analyse a text necessarily means working on and with the text itself. Therefore an

examination of advertising spots will also work with the spots and their content. Little else is

there than indeed to use content analysis. Using other methods would in this case need an

other objective of the study or they could be used to provide supplementary data to the study.

For example, when using a survey, this could be done to determine if people had seen any

political advertising spots, to find out which of them they remembered and what bits of them.

However, after nearly five years in the case of the 1998 elections and nearly one year after

2002, the results would quite possibly be extremely unreliable of the real viewing figures.

Apart from that, the problem of cost and effort needed to create and carry out such a survey

with reasonable reliability exceeds by far my resources in both time and money. Still, as such

surveys have probably been carried out anyway by polling agencies during the election

campaign, the work would be pointless. Existing results, however, can shed further light on

special aspects of the research.

Another possible research method could have been the focus group. Again this method would

not allow for answering the proposed research question, so that it would have to be asked

differently. As qualitative method the focus would here be on the construction of meaning

through the members of the group and how negative they perceive the spots would be. The

question would have to be altered accordingly.27 However, even if the question would be

different and the focus group method could have been employed, there are practical

limitations on that again in my case. Firstly, all the spots are obviously in German, which

means that the research would have to be conducted in Germany, which would be a financial

problem again. Secondly, with the financial resources available (aka none), it is hardly

feasible to assemble enough groups to get reasonably reliable and valid results.

As we have seen, there are other methods available to undertake research into negative

political advertising. But for the particular question of asking how much negative content

there is in these spots, content analysis is the method of choice.

27Obviously, this is not the way proper research would proceed like. It is clear that the method usually follows

the research question, and not the other way round.

10


11/17

But why is this question of any interest? Why shall we ask examine the content, and not the

perception? There are two answers to this question. First of all, negative political advertising

is part of the so-called Americanisation of election campaigns. It is said that this spreads

around the world and in more and more countries parties adopt their campaigning to these

techniques, which suggests that also negativity increases in political television spots. To

examine this hypothesis, it is necessary to examine the spots on the content level. Secondly,

there has already been a lot of research on content of political television spots in various

countries over various periods of time. This gives the opportunity to compare the results of

this research with others undertaken earlier. Also, even though there has been some research

in this direction in Germany, most of it has not been too thorough and has not focused on

negativity, either. And, anyway, the spots of the last two elections have not been examined

yet. To get an idea of what awaits one when doing such a research, I will outline the method

of content analysis now.

Many scholars have written about content analysis, and all of these accounts differ from each

other, but the main steps to take in the actual conduct are generally the same. I will not do a

literature review here, but concentrate on the main advantages and disadvantages of CA, the

usual procedure and comment on the parts of the CA that will be especially important for me.

The most cited definition for content analysis is by Berelson: Content analysis is a research

technique for the objective, systematic, and quantitative description of the manifest content of

communication.28 Other definitions go a similar way: CA is a research technique for the

systematic classification and description of communication content according to certain

usually predetermined categories.29 And: CA may be defined as a methodology by which

the researcher seeks to determine the manifest content of written, spoken or published

communication bysystematic, objective, and quantitative analysis.30 A little different is

Krippendorf: CA is a research technique for making replicable and valid inferences from

data to text.31 Whatever the exact definition, the key issues here are that communication

contentis analysed, that the analysis is a quantitative one, and that the analysis issystematic

and objective. Objectivity means not absolute truth here, but it means that the categories of

analysis [are] defined so clearly that different persons can apply them to the same content and

28 Berelson 1952, cited in Hansen et al. 1998, p. 94.29

Wright (1986), cited in Berger 2000, p. 173.30

Zito (1975), cited in Berger 2000, p. 174. Original emphasis.31

Krippendorf 1980, p. 21.

11


12/17

get the same results. () Objective means the results depend upon the procedure, not the

analyst.32 The systematic analysis is achieved by creating an analysis system that is applied

to the content in an automatic manner; the same way every time. Quantitative means the

translation of content into numerical values such as frequencies or percentages. The question

of content, however, seems to be the most disputed. While there are many advocates of the

view that only the manifestcontent, that is, the content as it appears rather than as the analyst

feels it 33, has to be coded, there are others who argue that in relying on the manifest content a

good proportion of meaning is lost, as a text is not only the sum of its components. Both sides

have rational arguments: Stempel argues that objectivity cannot be maintained when manifest

content is abandoned. By doing this, he claims, a subjective interpretation would come into

play that would not allow for high reliability and reproducibility. Krippendorf on the other

hand supports a high-inference system when doing content analysis. Analysing only manifest

content he finds not interesting. He sees that coding textual units into conceptual categories is

inevitably inferential; and even has to take the context of the texts into account to bear

meaningful results. Also Weber states: content-analytic procedures that restrict themselves to

themes that are stated explicitly would certainly important vindications.34 Obviously

high-inferential systems have greater problems demonstrating reliability; in view of the

greater realness of what can be counted they are, however, more desirable. Still, there needs

to be a clear-cut definition of coding categories, which will be explained later.

Another big criticism of content analysis is its concept of quantification and the interpretation

of this counting. As words and occurrences of words are translated into numbers like

percentages or frequencies, the big question here is what these numbers mean. There has been

a tendency earlier to make statements only from the frequency with which symbols occur.

However, it seems obvious that frequency alone cannot create meaning and does not

necessarily mean anything. It would clearly be nave to assume that a television serial

showing ten incidents of cigarette smoking is ten more times likely to influence viewers to

smoke than a television serial showing only one incident35 However, as Hansen explains,

that does not discredit CA itself, it only points to the need for placing what is counted in CA

within a theoretical framework which articulates, in the form of a model of communication

influence, the social significance and meaning of what is being counted.36 This is mirrored

32 Stempel 1989, p. 125.33 For example Berelson 1952, Stempel 1989 and Berger 2000.34

Weber 1990, p. 76.35

Hansen et al. 1998, p. 96.36

Hansen et al. 1998, p. 96.

12


13/17

by Krippendorf, who sees the need for an analytical construct. This construct provides the

rules for inferences about and to the context.37

The actual process of content analysis has been described as a series of different distinct steps,

whose numbers differs greatly from scholar to scholar. However, the main tasks remain the

same. I will stick to the steps suggested by Hansen et al., as they concentrate on the major

stages that cover the essence of the analysis. Furthermore I will not specify all possible

problems here in detail but concentrate on those especially relevant for my research.

The first question is obviously the definition of the research problem. This has been illustrated

in greater detail earlier in this essay. The second step is the selection of media and sample.

Usually, when doing CA, one is confronted with a huge body of material that has to be cut

down, especially for practical reasons. This can be done in two steps: First, it is necessary to

define clearly what body of media will be analysed, described and characterised. Next, it is

often desirable and necessary to choose a representative sample from this body of media

content.38 As with my topic, sampling is pretty straightforward. All television spots of the

five major parties during two election campaigns amount to approximately 30 minutes. To

analyse this body seems feasible. However, a certain sampling has been done here, which was

the decision to analyse only the spots of the major parties, as defined being those parties who

where represented in the Bundestag in the previous legislative period. Also the selection has

been made to look only at television spots and not at posters, advertisements in newspapers,

or radio spots. Also choosing the material of the last two elections can be numbered among

sampling.

The most challenging part of a CA is probably the definition of analytical categories. It is also

the most important one. However, this step actually consists of two different acts. The first

problem is to define the unit of analysis, that is, the basic unit of text to be classified.

According to Weber there are six possible options: Word, word sense, sentence, theme,

paragraph or whole text. Using the word as category has the advantage of high reliability,

although there might be problems of loosing meaning without the context. When interested in

references made to something, using the sentence as a unit is the best choice, as it provides

more contextual material. When using a whole text, though, detail might be lost when coding

as various themes can occur in a text and it cannot always be allocated an unambiguous

37Krippendorf 1980, p. 27.

38Hansen et al. 1998, p. 100.

13


14/17

reading to the text. For the research on political television spots, previous researchers have

acknowledged this problem: our method of dichotomising the sample into positive and

negative ads by determining a dominant focus is useful for analysis but may underestimate

the amount of negative information present even in a positive ad.39 In view of this the best

way to define my unit of analysis then might be the sentence.

When it comes to coding categories there are two general choices to make. Firstly, the

categories are to be mutually exclusive. If a recording unit can be classified simultaneously

in two or more categories then it is possible that the results are dubious.40 Secondly,

there is a decision to make how broad or narrow the categories are to be. It might be useful to

have narrow categories but assign these to different broader categories as to be able to get a

general overview. Also, when choosing the categories, one can either come up with an own

system or use already tested in other, similar research. When using existing systems, the

advantage is that one knows that the system is workable. Furthermore, it is much easier to

compare results and relate them to other studies when having used the same or similar

systems.41 For my research it seems useful to adopt an existing system, although some

adaptations will have to be made in order to match the particularities of German political

spots.

After having defined units and categories of analysis, one can basically start to perform the

analysis. However, two things remain which have to be thought about. The units and

categories have to be tested in two ways: Firstly, the coding has to be tried out in a small sub-

sample to reveal any possible inadequacies or inconsistencies in the system. Also there should

be a test about the reliability of the coding process. An inter-coder reliability test has to made

to assure that different coders do not come up with different findings, and an intra-coder test

has to be performed to see whether one particular coder still codes a text the same after some

time. If checks reveal considerable divergence then it is necessary to tighten up the

coding guidelines, to make the coding instructions and definitions clearer.42 For my project,

although I will be the only coder, I still plan to perform an inter-coder reliability test with at

least one person once. Intra-coder reliability will be checked as well, simply by coding the

first coded item again after a couple of days. Hopefully a reasonable reliability will develop.

39 Kaid & Johnston 1991, p. 63.40

Weber 1990, p. 23.41

Stempel 1989.42

Hansen et al. 1998, p. 121.

14


15/17

As my CA is rather simple and small-scale, the results can easily (hopefully) be analysed by

hand, without the need to use sophisticated computer programs. Analysis will refer to the

theoretical framework established although, as Hansen et al. note, it is also important to be

flexible and open-minded in the process of analysing the data.43 Sometimes new dimensions

appear which have not been thought of before.

The analysis will furthermore give more evidence about what problems this particular

methodology has and what issues might be outstanding with regard to the research field. It is,

however, possible to think of a couple of things where more time and money might have

broadened or changed the analytical framework. As explained earlier, CA as a quantitative

analysis method will not give any hints on how people actually perceive the spots. Even

though the CA itself might have found a high amount of negativity, in the textual context this

might not be clear to people and they might perceive it as completely different. Conducting

focus groups on this topic would be a good check on the validity of the findings as well as it

would shed further light on effectiveness of negative strategies in the spots. Another useful

thing might be not only to concentrate on the television spots but also broaden the research on

other campaigning tools, such as other advertising or press briefings and speeches. However,

given the limited time and resources, this is not feasible.

In conclusion some points can be made. Researchers should be aware of their ontological and

epistemological position to know how to properly embark on a research topic. Without this

prerequisite they are likely to get caught in contradictory statements and strategies without

knowing it. When this is accomplished, the methodology of the research project has to be

thought of and after thorough consideration of the advantages and problems of the strategy the

actual project can start. Ideally the conducted research will finally present itself as a coherent

whole and furthermore bear some new and interesting findings.

43Hansen et al. 1998, p. 122.

15


16/17

Bibliography:

Berger, Arthur Asa (2000):Media and Communication Research Methods: An Introduction to

Qualitative and Quantitative Approaches. Thousand Oaks, California: Sage Publications.

Bowers, John (1970): Content Analysis in Emmert, Philip, Brooks, William (eds.):Methods

of Research in Communication, Boston: Houghton, pp. 291-314.

Gavin, Neil T. (1998): Generalization, reliability and validity in media research in Gavin,

Neil T. (ed.): The Economy, Media and Public Knowledge. London: Leicester University

Press, pp. 171-174.

Hansen, Anders, Cottle, Simon, Negrine, Ralph, Newbold, Chris (1998):Mass

Communication Research Methods. New York: New York University Press.

Hay, Colin (2002):Political analysis: a critical introduction. Basingstoke: Palgrave.

Hollis, Martin and Smith, Steve (1991):Explaining and Understanding International

Relations. Oxford: Claredon Press.

Holtz-Bacha, Christina and Kaid, Lynda Lee (1995): Television Spots in German National

Elections in Kaid, Lynda Lee & Holtz-Bacha, Christina (eds.): Political Advertising in

Western Democracies. Parties and Candidates on Television. London: SAGE, pp. 61-88.

Johnston, Anne (1991): Political Broadcasts: An Analysis of Form, Content, and Style in

Presidential Communication in Kaid, Lynda Lee, Gerstl, Jacques, Sanders, Keith (eds.):

Mediated politics in two cultures. New York: Praeger, pp. 59-72.

Kaid, Lynda Lee and Johnston, Anne (1991): Negative versus Positive Television

Advertising in U.S. Presidential Campaigns, 1960-1988 inJournal of Communication 41(3),

pp. 53-64.

Krippendorf, Klaus: (1980): Content Analysis. An Introduction to Its Methodology. London:

Sage.

Lunt, Peter and Livingstone, Sonia (1996): Rethinking the Focus Group in Media and

Communications Research,Journal of Communication 46(2), pp. 79-98.

Marsh, David and Furlong, Edward (2002): Ontology and Epistemology in Political Science

in Marsh, David and Stoker, Gerry (eds.): Theory and Methods in Political Science, 2nd

edition. Basingstoke: Palgrave.

Read, Melvyn and Marsh, David (2002): Combining Quantitative and Qualitative Methods

in Marsh, David and Stoker, Gerry (eds.): Theory and Methods in Political Science, 2nd

edition. Basingstoke: Palgrave.

Robson, Colin (2002):Real World Research, 2nd edition. Oxford: Blackwell.

Sayer, Andrew (1984):Method in Social Science. A realist approach. London: Hutchinson.

16


17/17

Schmidt, Siegfried J. (1994a): Die Wirklichkeit des Beobachters [The reality of the

observer] in Merten, Klaus, Schmidt, Siegfried J. and Weischenberg, Siegfried (eds.):Die

Wirklichkeit der Medien. [The reality of the media]. Opladen: Westdeutscher Verlag, pp. 3-

19.

Schmidt, Siegfried J. (1994b): Konstruktivismus in der Medienforschung: Konzepte,Kritiken, Konsequenzen [Constructivism in media science: concepts, critiques,

consequences] in Merten, Klaus, Schmidt, Siegfried J. and Weischenberg, Siegfried (eds.):

Die Wirklichkeit der Medien. [The reality of the media]. Opladen: Westdeutscher Verlag, pp.

592-623.

Semetko, Holli and Schoenbach, Klaus (1995): The Media and the campaign in the new

Germany in Conradt, D., Kleinfeld, G., Romoser, G., Soe, Ch. (eds.): Germanys New

Politics. Oxford: Berghahn Books, pp. 61-80.

Stempel, Guido (1989): Content Analysis in Stempel, Guido H. & Westley, Bruce H. (eds.):

Research Methods in Mass Communication, 2nd edition. Prentice-Hall, Englewood Cliffs, pp.124-36.

17

Documents

Ont Epis Meth